摘要
随着虚拟专用网技术广泛运用,Open VPN流量分类识别在网络管理和空间安全扮演重要角色,由于基于传统加密流量识别技术在特征提取及选择方面成效不佳,提出一种基于卷积自编码的加密流量识别方法,将流量样本预处理为图片样本数据集IOVTD,集成特征提取与分类器功能,自动学习原始输入和预期输出之间的非线性关系,利用CAE的无监督特性结合多层感知机的有监督分类学习优势达成对Open VPN机密流量的识别分类。实验结果表明,数据集充足条件下,基于卷积自编码的加密流量识别方法各项性能指标均优于传统加密流量识别方法,总体识别精准率、召回率及F1-Score最高均可达99.87%以上。
With the wide use of virtual private network technology,traffic classification of Open VPN plays an important role in network management and cyberspace security.To solve the problem that traffic identification based on traditional methods is not effective in feature extraction and selection,this paper proposes an encrypted traffic identification method based on convolutional auto-encoding(CAE),which preprocesses traffic samples into IOVTD,integrates feature extraction and classifier functions,automatically learns the nonlinear relationship between the original input and the expected output,and utilizes the unsupervised nature of CAE and the supervised classification learning advantages of multi-layer perceptron to realize the identification and classification of Open VPN encrypted traffic.The experimental results show that the performance of the encrypted traffic identification method based on convolutional auto-encoding is better than the traditional one based on sufficient data set.The best overall recognition rate of accuracy,recall and F1-Score can reach more than 99.87%.
作者
郭路路
段明
王磊
刘鎏
李玎
GUO Lulu;DUAN Ming;WANG Lei;LIU Liu;LI Ding(Henan Key Laboratory of Network Cryptography, Zhengzhou 450001, China;State Key Laboratory of Mathematical Engineering and Advanced Computing, Zhengzhou 450001, China)
出处
《信息工程大学学报》
2019年第4期410-416,共7页
Journal of Information Engineering University
基金
国家重点研发计划资助项目(2016YFB0801505)。
